The study, published in the open-access journal Heliyon, was led by Asad Ullah and Junfeng Jiang, with co-authors Mashael Bakhit and Yanjie Li. Its central question is deceptively simple: does the circular economy actually cushion the environmental damage caused by industrialization, and does it make green innovation more effective? To answer it, the team moved beyond the average-based models that dominate the literature, which can mask enormous differences between, say, Sweden’s low-carbon industrial base and Poland’s coal-dependent manufacturing sector. Instead, they deployed a Panel Quantile Autoregressive Distributed Lag model with an error correction mechanism, a technique that estimates effects separately at different quantiles of the emissions distribution, from the cleanest to the most carbon-intensive economies.

The stakes are considerable. Energy-intensive industries such as steel, chemicals and manufacturing account for roughly 27 percent of industrial greenhouse gas emissions in Europe, along with more than 35 percent of sulfur oxide emissions and 31 percent of nitrogen oxide emissions. Industrialization remains a pillar of economic development and structural transformation across the bloc, yet it continues to drive carbon emissions, energy consumption and material extraction. The European Green Deal and its Circular Economy Action Plan were designed precisely to break this link, promoting recycling, reuse, waste reduction and resource recovery as routes to decoupling growth from environmental degradation. Circular activities are already estimated to generate more than €600 billion in added value through enhanced resource use and waste recovery, and about 11.8 percent of EU material use now comes from recycled and secondary materials.

To capture circularity in a single measure, the researchers built a composite index from five Eurostat indicators: the circular material use rate, municipal waste recycling rate, resource productivity, packaging waste recycling and the e-waste recycling rate. Principal component analysis confirmed the index was statistically robust, with the first component explaining 79.69 percent of the total variance and an overall Kaiser-Meyer-Olkin measure of 0.785, indicating good sampling adequacy. Environmental performance, the dependent variable, was measured as per capita carbon dioxide emissions from the World Development Indicators. Industrialization was proxied by industry value added as a share of GDP, while green innovation was captured through environment-related patent applications from the OECD database. Trade openness, renewable energy consumption and economic growth served as control variables.

The preliminary diagnostics alone tell an important story. Tests by Pesaran, Friedman and Frees all revealed significant cross-sectional dependence among the 27 economies, meaning that shocks in one country ripple through the others, as one would expect in a deeply integrated single market. Slope homogeneity tests rejected the assumption that all countries respond identically, validating the quantile approach. Second-generation unit root tests showed a mix of stationary and non-stationary series, justifying the ARDL framework, and cointegration tests confirmed a long-run equilibrium relationship among the variables. The error correction term was negative and significant at the 10th, 25th and 50th quantiles, indicating that lower and middle-emission economies correct roughly 7.7 percent of any short-run deviation each period. Strikingly, at the 75th and 90th quantiles the adjustment was statistically insignificant, suggesting that the highest-emitting countries face structural constraints, fossil fuel reliance and slower institutional change that prevent them from reverting to a decarbonization path.

The long-run results are where the study delivers its most provocative findings. Industrialization raised per capita carbon dioxide emissions at every quantile, confirming that the scale effect of industrial expansion persists across the bloc, though the magnitude varied with the emissions level. Green innovation showed a genuinely heterogeneous, non-linear pattern: it was associated with higher emissions at the 10th quantile, consistent with the innovation lag theory, in which research, development, infrastructure investment and commercialization temporarily demand more energy and materials before benefits materialize. It reduced emissions at the 25th quantile, but was linked to higher emissions at the median and 75th quantiles, possibly reflecting rebound effects or the coexistence of innovation with carbon-intensive industrial growth. At the very top of the distribution, the effect faded to statistical insignificance, implying that innovation alone cannot offset entrenched carbon dependence in the most polluting economies.

The circular economy’s direct effect was equally polarized. It was insignificant in the cleanest economies, dramatically negative in the lower half of the distribution, and yet positive at the median, before turning strongly negative again at the 75th and 90th quantiles, where coefficients reached −0.2001 and −0.2662. The positive median result may seem counterintuitive, but the authors point to the energy demands of the circular transition itself: collection systems, reverse logistics, recycling infrastructure and reprocessing all consume power, and if that power is not yet clean, early-stage circularity can temporarily raise emissions. The benefits grow strongest in high-emission, material-intensive economies, where the potential for resource efficiency gains is largest. This contradicts the common assumption that circular practices reduce emissions uniformly, regardless of implementation stage.

The moderation analysis, the heart of the paper, adds further nuance. The interaction between industrialization and circularity was negative and significant at the median quantile, meaning circular systems can genuinely mitigate the environmental toll of industrial growth when they improve resource use. But at the 25th quantile the interaction was slightly positive, and at the 75th and 90th quantiles it turned positive again, suggesting that rapid industrial expansion in highly industrialized economies may outpace the development of circular infrastructure. The interaction between green innovation and circularity, by contrast, was negative and significant at the 10th, 50th and 75th quantiles, indicating that the two strategies are complementary across much of the distribution: innovation makes circular production cleaner, while circularity helps innovation deliver its promised emission cuts. The synergy weakened at the extremes, with carbon lock-in likely blunting the innovation-circularity pairing among the very highest emitters. A robustness check using Driscoll-Kraay fixed-effects estimation, which addresses cross-sectional dependence, heteroskedasticity and serial correlation, largely confirmed these patterns, with the innovation-circularity interaction remaining significantly negative and the model explaining nearly 60 percent of within-country variation.

For the EU’s policymakers, the message is sobering but actionable. Renewable energy consumption emerged as the most reliable emission-reduction lever, significantly lowering emissions from the 10th through the 75th quantiles, though even it turned positive at the highest emission level, hinting at energy security constraints and grid limitations in the most carbon-intensive states. The authors argue that innovation policy must move beyond blanket subsidies toward conditional measures that account for where a technology sits in its life cycle, and that circular economy policy should not be treated as waste management but as industrial transformation, integrated with renewable energy deployment, low-carbon transport and recycling efficiency regulation. Lower-emission economies should focus on building circular infrastructure, intermediate emitters on decarbonizing circular activities, and the most polluting economies on coupling circularity with deep industrial upgrading. The study’s own caveats are worth heeding: per capita carbon dioxide captures only part of environmental quality, patents are an imperfect proxy for innovation in practice, and the estimates describe long-run associations rather than proven causal effects. Yet as the EU races toward its 2030 and 2050 targets, the research makes one thing unmistakably clear: there is no single dial to turn. Decarbonization in Europe will be won or lost country by country, quantile by quantile, depending on whether circularity, innovation and clean energy are deployed together, in the right places, at the right time.

Subject of Research: The moderating role of the circular economy in the relationship between green innovation, industrialization and environmental performance in EU-27 countries

Article Title: Circular economy as a moderator of the green innovation–industrialization nexus: evidence from environmental performance in EU-27 countries, 2000–2024

Article References: Circular economy as a moderator of the green innovation–industrialization nexus: evidence from environmental performance in EU-27 countries, 2000–2024. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: circular economy, green innovation, industrialization, carbon emissions, European Green Deal, EU-27, environmental performance, PQ-ARDL-ECM, renewable energy, recycling, climate neutrality, panel quantile analysis